The Wasted Times of Google (part 1)

Fast forward a century. When people discuss the groundbreaking companies of the 21st century that profoundly altered the course of human civilization, I believe Google will be among them.

Dating back to its inception in 1998, this tech powerhouse has instigated revolutions in various fields. Like a young calf facing a giant, Google dethroned the old order. From search to Gmail, from maps to online documents, from Android phones to the Chrome browser, from autonomous driving to the Transformer neural network architecture, from MapReduce to Spanner, from AlphaGo to AlphaFold - Google's technical revolutions are vast, its disruption ubiquitous, and its impact on human life profoundly far-reaching.

But as this once mighty oak has strived to reach the sky, around its unnoticed parts, toxic thorny vines have taken root and grown. One day, these vines wrapped around the tree, obscuring its grandeur, and people realized that the branches began to wither, and its leaves to fall. Google, once a place that captivated the idealistic young people no longer seems to be so, and its once unstoppable development has been thwarted in many fields.

What happened?

No change this dramatic happens overnight. The shifts in Google's landscape we see today can be traced back to vibes a decade ago.

Rewinding nineteen years.

In the spring of 2005, I left Microsoft to join Google. The company had recently made its debut on NASDAQ and exuded an optimistic atmosphere throughout.

Google, at that time, had over 6,000 employees, including more than 4,000 engineers. The Kirkland branch in Washington State, where I worked, housed about thirty employees.

In contrast, Microsoft was ten times larger, and the CEO at the time, Steve Ballmer, had just thrown a chair out of his window, sounding the trumpets to "bury Google". With AltaVista, Yahoo!, Ask, Lycos, InfoSeek, and MSN all coveting the throne of Google search, it was an intense game.

Yet despite an uncertain future, Google remained steadfast, bubbling with confidence.

Here's why. Google's ranks teemed with talent ready to face external threats, thriving in the heat of challenges.

Back then, Google had gained notoriety for how difficult it was to join. I referred many friends, classmates, former colleagues, and ex-bosses for interviews at Google. Among these highly competent individuals, less than a tenth managed to successfully navigate through the strenuous interview process.

Only when I was involved in recruiting did I understand that each interviewer had veto power during the six rounds of interviews. While this approach might lead to rejecting a thousand potentially good hires, it also ensured that the likelihood of recruiting mediocre talent was minimal.

Google's interview process was notoriously long — from the initial touchpoint to making an offer, it was not unusual to take two to three months. This was one cause for candidate discontent towards Google's recruitment. A colleague, who started on the same day as I, was so frustrated with Google's slow recruitment process that he initially accepted an offer from another company, only to switch to Google two months later once he received their offer.

In a bid to attract top engineers, Google once posted a math problem on a billboard along California's freeway 101. The answer, a ten-digit website address, was a direct line to Google's recruitment department if you solved the problem correctly.

You might see it as a gimmick, but this kind of geeky tactic undeniably deepens engineers' sense of belonging with the company. Birds of a feather flock together, as the old saying goes. Teaming up and working alongside the industry's best at Google became the aspiration of many passionate young individuals at the time.

A decade later, Google played this trick once more. When users searched for Python programming information, a Google Easter egg would reveal a link to a mysterious website. After solving a series of programming problems on the site, users would be invited to leave their contact details and would quickly receive an email from Google's recruitment department. Google knew that anyone capable of reaching this far was no ordinary.

Google did not let me down. After joining the company, I found myself engrossed in the internal email lists. In the days before Google+ (how many people remember this once highly anticipated product?), these threads were a battleground where the experts showed off their skills. They worked on different projects but would often discuss and exchange views on technical problems outside of their projects.

Being a tech-curious engineer, reading these emails daily felt as thrilling as attending a concert. Engineers like Jeff Dean, who, as the legend goes, at one time wrote 30% of the company's code with his partner Sanjay Ghemawat, held rockstar status within Google.

I didn't merely observe these threads; I participated in many discussions. Showing your weaknesses in front of experts can be a great opportunity for personal growth.

Whenever I sang Google's praises to my friends, I would repeatedly mention one crucial thing: Google's most considerable benefit isn't the generous salary, stocks, gourmet food, or massages—it's the company's concentration of talent. “Better to have fewer but excellent, than to have more but mediocre” was once a Google creed that was executed earnestly throughout the first decade of the 21st century. The result was that Google managed to maintain its geek paradise engineering culture while experiencing rapid business growth.

The abundance of memes about bad team players online showed everyone's reluctance to tolerate underperforming colleagues. When top talents meet, their combined energy could spark the company's next growth explosion. Netflix founder Reed Hastings once said that A-players only want to work with other A-players, thus his company will pay top dollar to attract the best talent. It is crucial to ensure that recruitment standards do not slip because once the company starts hiring B-players, they will attract C-players and so on, subsequently ruining the company culture.

Unfortunately, Google has been heading down the path warned by Hastings these past ten years.

In 2014, Google's employee count surpassed 50,000. By the time I left Google in 2022, the parent company Alphabet's employee count had reached a whopping 190,000. With such a massive company expanding so rapidly, can we really believe that the hiring standards haven't been lowered?

The number of top talents in the market is finite and won’t change suddenly. To increase the workforce by 140,000 employees in just a few years... it's simply not feasible to expect top talents to magically appear. As a result, the average competence level of the employees visibly dropped. This does not imply that subsequent hires were unqualified, but rather the unqualified rate increased.

The rise of platforms like LeetCode had made it easier to enhance one’s interview performance, potentially burying real gems. This certainly had increased the challenge of talent acquisition for the company.

A key issue with persistent low standards is that over time everyone adjusts, thinking it normal. Memories of the company's initial glory days filled with heroic figures become increasingly remote, only recalled occasionally by old timers. Many colleagues who joined the company after the competence decline never had the opportunity to experience working side by side with top talents.

According to the Exclusion Principle of talents, an increase in mediocre teammates leads to the exit of top talents. They will seek environments with higher talent density.

In the high-tech industry, a headcount war is not the winning strategy. Redundant employees are a burden to the company, especially during layoffs. OpenAI, with fewer than a thousand employees, caught Google off guard. At its breakout point, Midjourney only had 11 full-time employees.

Perhaps we should all learn from past mistakes: Never fall into the trap of competing for mediocre talent out of fear of missing out.

(Continue to part 2)